Refactor `get_tau_sigma` and support lists of variables

[tvwenger]

What is this PR about? This PR refactors the get_tau_sigma logic in order to support lists of Variables and fixes https://github.com/pymc-devs/pymc/issues/6987.

N.B. The original code had a positivity check for non-Variable arguments, but not for Variable arguments. Since this PR casts the arguments to Variable, I've removed the positivity check for non-Variable arguments.

Checklist

• [X] Explain important implementation details 👆
• [X] Make sure that the pre-commit linting/style checks pass.
• [X] Link relevant issues (preferably in nice commit messages)
• [X] Are the changes covered by tests and docstrings?
• [X] Fill out the short summary sections 👇

Major / Breaking Changes

• get_tau_sigma now always returns a Variable, even when scalars are passed

New features

• N/A

Bugfixes

• https://github.com/pymc-devs/pymc/issues/6987

Documentation

• N/A

Maintenance

• Updated tests

---

:books: Documentation preview :books:: https://pymc--6988.org.readthedocs.build/en/6988/

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 90.00000% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 77.73%. Comparing base (46f1675) to head (bad3b80). Report is 5 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #6988       +/-   ##
===========================================
- Coverage   89.60%   77.73%   -11.88%     
===========================================
  Files         100      100               
  Lines       16888    16880        -8     
===========================================
- Hits        15133    13121     -2012     
- Misses       1755     3759     +2004     

Files	Coverage Δ
pymc/distributions/shape_utils.py	59.89% <ø> (-36.05%)	:arrow_down:
pymc/model/core.py	72.15% <ø> (-19.18%)	:arrow_down:
pymc/model_graph.py	14.57% <ø> (-62.82%)	:arrow_down:
pymc/pytensorf.py	73.61% <ø> (-17.42%)	:arrow_down:
pymc/distributions/continuous.py	64.64% <90.00%> (-32.90%)	:arrow_down:

... and 40 files with indirect coverage changes

[ricardoV94]

@tvwenger would you like to update this PR? Would only need to add some regression tests for the original issue

[tvwenger]

@ricardoV94 Thanks for the reminder and sorry for the delay. I added some regression tests, so this should be good to go once the checks are finished.

[tvwenger]

@ricardoV94 This PR is ready for review

[tvwenger]

@ricardoV94 Ready for review when checks are done